Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
40.71% |
103 / 253 |
|
25.81% |
8 / 31 |
CRAP | |
0.00% |
0 / 1 |
| Hooks | |
40.71% |
103 / 253 |
|
25.81% |
8 / 31 |
1975.87 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| onRegistration | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
6 | |||
| onBeforeInitialize | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| onApiBeforeMain | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| initializeForRequest | |
95.24% |
20 / 21 |
|
0.00% |
0 / 1 |
4 | |||
| overrideNumeric | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| overrideMinimumShouldMatch | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
3 | |||
| overrideSecret | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
3 | |||
| overrideYesNo | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| overrideMoreLikeThisOptionsFromMessage | |
25.00% |
11 / 44 |
|
0.00% |
0 / 1 |
226.19 | |||
| isMinimumShouldMatch | |
62.50% |
5 / 8 |
|
0.00% |
0 / 1 |
9.58 | |||
| overrideMoreLikeThisOptions | |
100.00% |
19 / 19 |
|
100.00% |
1 / 1 |
2 | |||
| onSoftwareInfo | |
0.00% |
0 / 15 |
|
0.00% |
0 / 1 |
12 | |||
| onSpecialSearchResultsAppend | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
12 | |||
| addSearchFeedbackLink | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
2 | |||
| onPrefixSearchExtractNamespace | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
6 | |||
| prefixSearchExtractNamespaceWithConnection | |
53.33% |
8 / 15 |
|
0.00% |
0 / 1 |
5.63 | |||
| onSearchGetNearMatch | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| handleSearchGetNearMatch | |
0.00% |
0 / 22 |
|
0.00% |
0 / 1 |
42 | |||
| onResourceLoaderGetConfigVars | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| getConfig | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| getConnection | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| onShowSearchHitTitle | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
12 | |||
| onAPIAfterExecute | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
20 | |||
| onSpecialSearchResults | |
0.00% |
0 / 17 |
|
0.00% |
0 / 1 |
20 | |||
| addWordCount | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
12 | |||
| onGetPreferences | |
100.00% |
13 / 13 |
|
100.00% |
1 / 1 |
3 | |||
| autoCompleteOptionsForPreferences | |
100.00% |
19 / 19 |
|
100.00% |
1 / 1 |
4 | |||
| onUserGetDefaultOptions | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| onSpecialStatsAddExtra | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| onAPIQuerySiteInfoStatisticsInfo | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace CirrusSearch; |
| 4 | |
| 5 | use CirrusSearch\Search\FancyTitleResultsType; |
| 6 | use HtmlArmor; |
| 7 | use ISearchResultSet; |
| 8 | use MediaWiki\Actions\ActionEntryPoint; |
| 9 | use MediaWiki\Api\ApiBase; |
| 10 | use MediaWiki\Api\ApiMain; |
| 11 | use MediaWiki\Api\ApiOpenSearch; |
| 12 | use MediaWiki\Api\Hook\APIAfterExecuteHook; |
| 13 | use MediaWiki\Api\Hook\APIQuerySiteInfoStatisticsInfoHook; |
| 14 | use MediaWiki\Config\Config; |
| 15 | use MediaWiki\Config\ConfigFactory; |
| 16 | use MediaWiki\Context\RequestContext; |
| 17 | use MediaWiki\Hook\ApiBeforeMainHook; |
| 18 | use MediaWiki\Hook\BeforeInitializeHook; |
| 19 | use MediaWiki\Hook\SoftwareInfoHook; |
| 20 | use MediaWiki\Hook\SpecialSearchResultsAppendHook; |
| 21 | use MediaWiki\Hook\SpecialSearchResultsHook; |
| 22 | use MediaWiki\Hook\SpecialStatsAddExtraHook; |
| 23 | use MediaWiki\Html\Html; |
| 24 | use MediaWiki\Logger\LoggerFactory; |
| 25 | use MediaWiki\MediaWikiServices; |
| 26 | use MediaWiki\Output\OutputPage; |
| 27 | use MediaWiki\Preferences\Hook\GetPreferencesHook; |
| 28 | use MediaWiki\Request\WebRequest; |
| 29 | use MediaWiki\ResourceLoader\Hook\ResourceLoaderGetConfigVarsHook; |
| 30 | use MediaWiki\Search\Hook\PrefixSearchExtractNamespaceHook; |
| 31 | use MediaWiki\Search\Hook\SearchGetNearMatchHook; |
| 32 | use MediaWiki\Search\Hook\ShowSearchHitTitleHook; |
| 33 | use MediaWiki\Specials\SpecialSearch; |
| 34 | use MediaWiki\Title\Title; |
| 35 | use MediaWiki\User\Hook\UserGetDefaultOptionsHook; |
| 36 | use MediaWiki\User\User; |
| 37 | use SearchResult; |
| 38 | |
| 39 | /** |
| 40 | * All CirrusSearch's external hooks. |
| 41 | * |
| 42 | * This program is free software; you can redistribute it and/or modify |
| 43 | * it under the terms of the GNU General Public License as published by |
| 44 | * the Free Software Foundation; either version 2 of the License, or |
| 45 | * (at your option) any later version. |
| 46 | * |
| 47 | * This program is distributed in the hope that it will be useful, |
| 48 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 49 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 50 | * GNU General Public License for more details. |
| 51 | * |
| 52 | * You should have received a copy of the GNU General Public License along |
| 53 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 54 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 55 | * http://www.gnu.org/copyleft/gpl.html |
| 56 | */ |
| 57 | class Hooks implements |
| 58 | UserGetDefaultOptionsHook, |
| 59 | GetPreferencesHook, |
| 60 | APIAfterExecuteHook, |
| 61 | ApiBeforeMainHook, |
| 62 | APIQuerySiteInfoStatisticsInfoHook, |
| 63 | BeforeInitializeHook, |
| 64 | PrefixSearchExtractNamespaceHook, |
| 65 | ResourceLoaderGetConfigVarsHook, |
| 66 | SearchGetNearMatchHook, |
| 67 | ShowSearchHitTitleHook, |
| 68 | SoftwareInfoHook, |
| 69 | SpecialSearchResultsHook, |
| 70 | SpecialSearchResultsAppendHook, |
| 71 | SpecialStatsAddExtraHook |
| 72 | { |
| 73 | /** @var ConfigFactory */ |
| 74 | private $configFactory; |
| 75 | |
| 76 | /** |
| 77 | * @param ConfigFactory $configFactory |
| 78 | */ |
| 79 | public function __construct( ConfigFactory $configFactory ) { |
| 80 | $this->configFactory = $configFactory; |
| 81 | } |
| 82 | |
| 83 | /** |
| 84 | * Extension registration callback (https://www.mediawiki.org/wiki/Manual:Extension.json/Schema#callback), called |
| 85 | * early on in the MW setup. |
| 86 | */ |
| 87 | public static function onRegistration(): void { |
| 88 | global $wgCirrusSearchClusters, $wgCirrusSearchWriteClusters; |
| 89 | if ( defined( 'MW_QUIBBLE_CI' ) ) { |
| 90 | // If running in quibble, disable writes: they will fail anyway and potentially lead to timeouts (T389895). |
| 91 | $wgCirrusSearchClusters = [ |
| 92 | 'default' => [ 'localhost' ], |
| 93 | ]; |
| 94 | $wgCirrusSearchWriteClusters = []; |
| 95 | } |
| 96 | } |
| 97 | |
| 98 | /** |
| 99 | * Hooked to call initialize after the user is set up. |
| 100 | * |
| 101 | * @param Title $title |
| 102 | * @param null $unused |
| 103 | * @param OutputPage $outputPage |
| 104 | * @param User $user |
| 105 | * @param WebRequest $request |
| 106 | * @param ActionEntryPoint $mediaWiki |
| 107 | */ |
| 108 | public function onBeforeInitialize( $title, $unused, $outputPage, $user, $request, $mediaWiki ) { |
| 109 | self::initializeForRequest( $request ); |
| 110 | } |
| 111 | |
| 112 | /** |
| 113 | * Hooked to call initialize after the user is set up. |
| 114 | * @param ApiMain &$apiMain The ApiMain instance being used |
| 115 | */ |
| 116 | public function onApiBeforeMain( &$apiMain ) { |
| 117 | self::initializeForRequest( $apiMain->getRequest() ); |
| 118 | } |
| 119 | |
| 120 | /** |
| 121 | * Initializes the portions of Cirrus that require the $request to be fully initialized |
| 122 | * |
| 123 | * @param WebRequest $request |
| 124 | */ |
| 125 | public static function initializeForRequest( WebRequest $request ) { |
| 126 | global $wgCirrusSearchPhraseRescoreWindowSize, |
| 127 | $wgCirrusSearchFunctionRescoreWindowSize, |
| 128 | $wgCirrusSearchFragmentSize, |
| 129 | $wgCirrusSearchPhraseRescoreBoost, |
| 130 | $wgCirrusSearchPhraseSlop, |
| 131 | $wgCirrusSearchLogElasticRequests, |
| 132 | $wgCirrusSearchLogElasticRequestsSecret, |
| 133 | $wgCirrusSearchEnableAltLanguage, |
| 134 | $wgCirrusSearchUseCompletionSuggester, |
| 135 | $wgCirrusSearchMustTrackTotalHits; |
| 136 | |
| 137 | if ( $request->getCheck( 'cirrusApproxTotalHits' ) ) { |
| 138 | $wgCirrusSearchMustTrackTotalHits = [ 'default' => false ]; |
| 139 | } |
| 140 | self::overrideMoreLikeThisOptionsFromMessage(); |
| 141 | |
| 142 | self::overrideNumeric( $wgCirrusSearchPhraseRescoreWindowSize, |
| 143 | $request, 'cirrusPhraseWindow', 10000 ); |
| 144 | self::overrideNumeric( $wgCirrusSearchPhraseRescoreBoost, |
| 145 | $request, 'cirrusPhraseBoost' ); |
| 146 | self::overrideNumeric( $wgCirrusSearchPhraseSlop[ 'boost' ], |
| 147 | $request, 'cirrusPhraseSlop', 10 ); |
| 148 | self::overrideNumeric( $wgCirrusSearchFunctionRescoreWindowSize, |
| 149 | $request, 'cirrusFunctionWindow', 10000 ); |
| 150 | self::overrideNumeric( $wgCirrusSearchFragmentSize, |
| 151 | $request, 'cirrusFragmentSize', 1000 ); |
| 152 | if ( $wgCirrusSearchUseCompletionSuggester === 'yes' || $wgCirrusSearchUseCompletionSuggester === true ) { |
| 153 | // Only allow disabling the completion suggester, enabling it from request params might cause failures |
| 154 | // as the index might not be present. |
| 155 | self::overrideYesNo( $wgCirrusSearchUseCompletionSuggester, |
| 156 | $request, 'cirrusUseCompletionSuggester' ); |
| 157 | } |
| 158 | self::overrideMoreLikeThisOptions( $request ); |
| 159 | self::overrideSecret( $wgCirrusSearchLogElasticRequests, |
| 160 | $wgCirrusSearchLogElasticRequestsSecret, $request, 'cirrusLogElasticRequests', false ); |
| 161 | self::overrideYesNo( $wgCirrusSearchEnableAltLanguage, |
| 162 | $request, 'cirrusAltLanguage' ); |
| 163 | } |
| 164 | |
| 165 | /** |
| 166 | * Set $dest to the numeric value from $request->getVal( $name ) if it is <= $limit |
| 167 | * or => $limit if upperLimit is false. |
| 168 | * |
| 169 | * @param mixed &$dest |
| 170 | * @param WebRequest $request |
| 171 | * @param string $name |
| 172 | * @param int|null $limit |
| 173 | * @param bool $upperLimit |
| 174 | */ |
| 175 | private static function overrideNumeric( |
| 176 | &$dest, |
| 177 | WebRequest $request, |
| 178 | $name, |
| 179 | $limit = null, |
| 180 | $upperLimit = true |
| 181 | ) { |
| 182 | Util::overrideNumeric( $dest, $request, $name, $limit, $upperLimit ); |
| 183 | } |
| 184 | |
| 185 | /** |
| 186 | * @param mixed &$dest |
| 187 | * @param WebRequest $request |
| 188 | * @param string $name |
| 189 | */ |
| 190 | private static function overrideMinimumShouldMatch( &$dest, WebRequest $request, $name ) { |
| 191 | $val = $request->getVal( $name ); |
| 192 | if ( $val !== null && self::isMinimumShouldMatch( $val ) ) { |
| 193 | $dest = $val; |
| 194 | } |
| 195 | } |
| 196 | |
| 197 | /** |
| 198 | * Set $dest to $value when $request->getVal( $name ) contains $secret |
| 199 | * |
| 200 | * @param mixed &$dest |
| 201 | * @param string $secret |
| 202 | * @param WebRequest $request |
| 203 | * @param string $name |
| 204 | * @param mixed $value |
| 205 | */ |
| 206 | private static function overrideSecret( &$dest, $secret, WebRequest $request, $name, $value = true ) { |
| 207 | if ( $secret && $secret === $request->getVal( $name ) ) { |
| 208 | $dest = $value; |
| 209 | } |
| 210 | } |
| 211 | |
| 212 | /** |
| 213 | * Set $dest to the true/false from $request->getVal( $name ) if yes/no. |
| 214 | * |
| 215 | * @param mixed &$dest |
| 216 | * @param WebRequest $request |
| 217 | * @param string $name |
| 218 | */ |
| 219 | private static function overrideYesNo( &$dest, WebRequest $request, $name ) { |
| 220 | Util::overrideYesNo( $dest, $request, $name ); |
| 221 | } |
| 222 | |
| 223 | /** |
| 224 | * Extract more like this settings from the i18n message cirrussearch-morelikethis-settings |
| 225 | */ |
| 226 | private static function overrideMoreLikeThisOptionsFromMessage() { |
| 227 | global $wgCirrusSearchMoreLikeThisConfig, |
| 228 | $wgCirrusSearchMoreLikeThisAllowedFields, |
| 229 | $wgCirrusSearchMoreLikeThisMaxQueryTermsLimit, |
| 230 | $wgCirrusSearchMoreLikeThisFields; |
| 231 | |
| 232 | $cache = MediaWikiServices::getInstance()->getLocalServerObjectCache(); |
| 233 | $lines = $cache->getWithSetCallback( |
| 234 | $cache->makeKey( 'cirrussearch-morelikethis-settings' ), |
| 235 | 600, |
| 236 | static function () { |
| 237 | $source = wfMessage( 'cirrussearch-morelikethis-settings' )->inContentLanguage(); |
| 238 | if ( $source->isDisabled() ) { |
| 239 | return []; |
| 240 | } |
| 241 | return Util::parseSettingsInMessage( $source->plain() ); |
| 242 | } |
| 243 | ); |
| 244 | |
| 245 | foreach ( $lines as $line ) { |
| 246 | if ( strpos( $line, ':' ) === false ) { |
| 247 | continue; |
| 248 | } |
| 249 | [ $k, $v ] = explode( ':', $line, 2 ); |
| 250 | switch ( $k ) { |
| 251 | case 'min_doc_freq': |
| 252 | case 'max_doc_freq': |
| 253 | case 'max_query_terms': |
| 254 | case 'min_term_freq': |
| 255 | case 'min_word_length': |
| 256 | case 'max_word_length': |
| 257 | if ( is_numeric( $v ) && $v >= 0 ) { |
| 258 | $wgCirrusSearchMoreLikeThisConfig[$k] = intval( $v ); |
| 259 | } elseif ( $v === 'null' ) { |
| 260 | unset( $wgCirrusSearchMoreLikeThisConfig[$k] ); |
| 261 | } |
| 262 | break; |
| 263 | case 'percent_terms_to_match': |
| 264 | // @deprecated Use minimum_should_match now |
| 265 | $k = 'minimum_should_match'; |
| 266 | if ( is_numeric( $v ) && $v > 0 && $v <= 1 ) { |
| 267 | $v = ( (int)( (float)$v * 100 ) ) . '%'; |
| 268 | } else { |
| 269 | break; |
| 270 | } |
| 271 | // intentional fall-through |
| 272 | case 'minimum_should_match': |
| 273 | if ( self::isMinimumShouldMatch( $v ) ) { |
| 274 | $wgCirrusSearchMoreLikeThisConfig[$k] = $v; |
| 275 | } elseif ( $v === 'null' ) { |
| 276 | unset( $wgCirrusSearchMoreLikeThisConfig[$k] ); |
| 277 | } |
| 278 | break; |
| 279 | case 'fields': |
| 280 | $wgCirrusSearchMoreLikeThisFields = array_intersect( |
| 281 | array_map( 'trim', explode( ',', $v ) ), |
| 282 | $wgCirrusSearchMoreLikeThisAllowedFields ); |
| 283 | break; |
| 284 | } |
| 285 | // @phan-suppress-next-line PhanTypePossiblyInvalidDimOffset |
| 286 | if ( $wgCirrusSearchMoreLikeThisConfig['max_query_terms'] > $wgCirrusSearchMoreLikeThisMaxQueryTermsLimit ) { |
| 287 | $wgCirrusSearchMoreLikeThisConfig['max_query_terms'] = $wgCirrusSearchMoreLikeThisMaxQueryTermsLimit; |
| 288 | } |
| 289 | } |
| 290 | } |
| 291 | |
| 292 | /** |
| 293 | * @param string $v The value to check |
| 294 | * @return bool True if $v is an integer percentage in the domain -100 <= $v <= 100, $v != 0 |
| 295 | * @todo minimum_should_match also supports combinations (3<90%) and multiple combinations |
| 296 | */ |
| 297 | private static function isMinimumShouldMatch( string $v ) { |
| 298 | // specific integer count > 0 |
| 299 | if ( ctype_digit( $v ) && $v != 0 ) { |
| 300 | return true; |
| 301 | } |
| 302 | // percentage 0 < x <= 100 |
| 303 | if ( !str_ends_with( $v, '%' ) ) { |
| 304 | return false; |
| 305 | } |
| 306 | $v = substr( $v, 0, -1 ); |
| 307 | if ( str_starts_with( $v, '-' ) ) { |
| 308 | $v = substr( $v, 1 ); |
| 309 | } |
| 310 | return ctype_digit( $v ) && $v > 0 && $v <= 100; |
| 311 | } |
| 312 | |
| 313 | /** |
| 314 | * Override more like this settings from request URI parameters |
| 315 | * |
| 316 | * @param WebRequest $request |
| 317 | */ |
| 318 | private static function overrideMoreLikeThisOptions( WebRequest $request ) { |
| 319 | global $wgCirrusSearchMoreLikeThisConfig, |
| 320 | $wgCirrusSearchMoreLikeThisAllowedFields, |
| 321 | $wgCirrusSearchMoreLikeThisMaxQueryTermsLimit, |
| 322 | $wgCirrusSearchMoreLikeThisFields; |
| 323 | |
| 324 | self::overrideNumeric( $wgCirrusSearchMoreLikeThisConfig['min_doc_freq'], |
| 325 | $request, 'cirrusMltMinDocFreq' ); |
| 326 | self::overrideNumeric( $wgCirrusSearchMoreLikeThisConfig['max_doc_freq'], |
| 327 | $request, 'cirrusMltMaxDocFreq' ); |
| 328 | // @phan-suppress-next-line PhanTypePossiblyInvalidDimOffset |
| 329 | self::overrideNumeric( $wgCirrusSearchMoreLikeThisConfig['max_query_terms'], |
| 330 | $request, 'cirrusMltMaxQueryTerms', $wgCirrusSearchMoreLikeThisMaxQueryTermsLimit ); |
| 331 | self::overrideNumeric( $wgCirrusSearchMoreLikeThisConfig['min_term_freq'], |
| 332 | $request, 'cirrusMltMinTermFreq' ); |
| 333 | self::overrideMinimumShouldMatch( $wgCirrusSearchMoreLikeThisConfig['minimum_should_match'], |
| 334 | $request, 'cirrusMltMinimumShouldMatch' ); |
| 335 | self::overrideNumeric( $wgCirrusSearchMoreLikeThisConfig['min_word_length'], |
| 336 | $request, 'cirrusMltMinWordLength' ); |
| 337 | self::overrideNumeric( $wgCirrusSearchMoreLikeThisConfig['max_word_length'], |
| 338 | $request, 'cirrusMltMaxWordLength' ); |
| 339 | $fields = $request->getVal( 'cirrusMltFields' ); |
| 340 | if ( $fields !== null ) { |
| 341 | $wgCirrusSearchMoreLikeThisFields = array_intersect( |
| 342 | array_map( 'trim', explode( ',', $fields ) ), |
| 343 | $wgCirrusSearchMoreLikeThisAllowedFields ); |
| 344 | } |
| 345 | } |
| 346 | |
| 347 | /** |
| 348 | * Hook called to include Elasticsearch version info on Special:Version |
| 349 | * @param array &$software Array of wikitext and version numbers |
| 350 | */ |
| 351 | public function onSoftwareInfo( &$software ) { |
| 352 | $version = new Version( self::getConnection() ); |
| 353 | $status = $version->get(); |
| 354 | // We've already logged if this isn't ok and there is no need to warn the user on this page. |
| 355 | if ( $status->isOK() ) { |
| 356 | $distributions = [ |
| 357 | 'elasticsearch' => '[https://www.elastic.co/elasticsearch Elasticsearch]', |
| 358 | 'opensearch' => '[https://opensearch.org OpenSearch]', |
| 359 | ]; |
| 360 | $value = $status->getValue(); |
| 361 | $desc = $distributions[$value['distribution']] ?? null; |
| 362 | if ( $desc === null ) { |
| 363 | LoggerFactory::getInstance( 'CirrusSearch' )->warning( |
| 364 | 'Unexpected software distribution [{name}] returned by Version check', |
| 365 | [ 'name' => $value['distribution'] ] |
| 366 | ); |
| 367 | } else { |
| 368 | $software[$desc] = $value['version']; |
| 369 | } |
| 370 | } |
| 371 | } |
| 372 | |
| 373 | /** |
| 374 | * @param SpecialSearch $specialSearch |
| 375 | * @param OutputPage $out |
| 376 | * @param string $term |
| 377 | */ |
| 378 | public function onSpecialSearchResultsAppend( $specialSearch, $out, $term ) { |
| 379 | $feedbackLink = $out->getConfig()->get( 'CirrusSearchFeedbackLink' ); |
| 380 | |
| 381 | if ( $feedbackLink ) { |
| 382 | self::addSearchFeedbackLink( $feedbackLink, $specialSearch, $out ); |
| 383 | } |
| 384 | |
| 385 | // Embed metrics if this was a Cirrus page |
| 386 | $engine = $specialSearch->getSearchEngine(); |
| 387 | if ( $engine instanceof CirrusSearch ) { |
| 388 | $out->addJsConfigVars( $engine->getLastSearchMetrics() ); |
| 389 | } |
| 390 | } |
| 391 | |
| 392 | /** |
| 393 | * @param string $link |
| 394 | * @param SpecialSearch $specialSearch |
| 395 | * @param OutputPage $out |
| 396 | */ |
| 397 | private static function addSearchFeedbackLink( $link, SpecialSearch $specialSearch, OutputPage $out ) { |
| 398 | $anchor = Html::element( |
| 399 | 'a', |
| 400 | [ 'href' => $link ], |
| 401 | $specialSearch->msg( 'cirrussearch-give-feedback' )->text() |
| 402 | ); |
| 403 | $block = Html::rawElement( 'div', [], $anchor ); |
| 404 | $out->addHTML( $block ); |
| 405 | } |
| 406 | |
| 407 | /** |
| 408 | * Extract namespaces from query string. |
| 409 | * @param array &$namespaces |
| 410 | * @param string &$search |
| 411 | * @return bool |
| 412 | */ |
| 413 | public function onPrefixSearchExtractNamespace( &$namespaces, &$search ) { |
| 414 | global $wgSearchType; |
| 415 | if ( $wgSearchType !== 'CirrusSearch' ) { |
| 416 | return true; |
| 417 | } |
| 418 | return self::prefixSearchExtractNamespaceWithConnection( self::getConnection(), $namespaces, $search ); |
| 419 | } |
| 420 | |
| 421 | /** |
| 422 | * @param Connection $connection |
| 423 | * @param array &$namespaces |
| 424 | * @param string &$search |
| 425 | * @return false |
| 426 | */ |
| 427 | public static function prefixSearchExtractNamespaceWithConnection( |
| 428 | Connection $connection, |
| 429 | &$namespaces, |
| 430 | &$search |
| 431 | ) { |
| 432 | $method = $connection->getConfig()->get( 'CirrusSearchNamespaceResolutionMethod' ); |
| 433 | if ( $method === 'elastic' ) { |
| 434 | $searcher = |
| 435 | new Searcher( $connection, 0, 1, $connection->getConfig(), $namespaces ); |
| 436 | $searcher->updateNamespacesFromQuery( $search ); |
| 437 | $namespaces = $searcher->getSearchContext()->getNamespaces(); |
| 438 | } else { |
| 439 | $colon = strpos( $search, ':' ); |
| 440 | if ( $colon === false ) { |
| 441 | return false; |
| 442 | } |
| 443 | $namespaceName = substr( $search, 0, $colon ); |
| 444 | $ns = Util::identifyNamespace( $namespaceName, $method ); |
| 445 | if ( $ns !== false ) { |
| 446 | $namespaces = [ $ns ]; |
| 447 | $search = substr( $search, $colon + 1 ); |
| 448 | } |
| 449 | } |
| 450 | |
| 451 | return false; |
| 452 | } |
| 453 | |
| 454 | /** @inheritDoc */ |
| 455 | public function onSearchGetNearMatch( $term, &$titleResult ) { |
| 456 | return self::handleSearchGetNearMatch( $term, $titleResult ); |
| 457 | } |
| 458 | |
| 459 | /** |
| 460 | * Let Elasticsearch take a crack at getting near matches once mediawiki has tried all kinds of variants. |
| 461 | * @param string $term the original search term and all language variants |
| 462 | * @param null|Title &$titleResult resulting match. A Title if we found something, unchanged otherwise. |
| 463 | * @return bool return false if we find something, true otherwise so mediawiki can try its default behavior |
| 464 | */ |
| 465 | public static function handleSearchGetNearMatch( $term, &$titleResult ) { |
| 466 | global $wgSearchType; |
| 467 | if ( $wgSearchType !== 'CirrusSearch' ) { |
| 468 | return true; |
| 469 | } |
| 470 | |
| 471 | $title = Title::newFromText( $term ); |
| 472 | if ( $title === null ) { |
| 473 | return false; |
| 474 | } |
| 475 | |
| 476 | $user = RequestContext::getMain()->getUser(); |
| 477 | // Ask for the first 50 results we see. If there are more than that too bad. |
| 478 | $searcher = new Searcher( |
| 479 | self::getConnection(), 0, 50, self::getConfig(), [ $title->getNamespace() ], $user ); |
| 480 | if ( $title->getNamespace() === NS_MAIN ) { |
| 481 | $searcher->updateNamespacesFromQuery( $term ); |
| 482 | } else { |
| 483 | $term = $title->getText(); |
| 484 | } |
| 485 | $searcher->setResultsType( new FancyTitleResultsType( 'near_match' ) ); |
| 486 | $status = $searcher->nearMatchTitleSearch( $term ); |
| 487 | // There is no way to send errors or warnings back to the caller here so we have to make do with |
| 488 | // only sending results back if there are results and relying on the logging done at the status |
| 489 | // construction site to log errors. |
| 490 | if ( !$status->isOK() ) { |
| 491 | return true; |
| 492 | } |
| 493 | |
| 494 | $contLang = MediaWikiServices::getInstance()->getContentLanguage(); |
| 495 | $picker = new NearMatchPicker( $contLang, $term, $status->getValue() ); |
| 496 | $best = $picker->pickBest(); |
| 497 | if ( $best ) { |
| 498 | $titleResult = $best; |
| 499 | return false; |
| 500 | } |
| 501 | // Didn't find a result so let MediaWiki have a crack at it. |
| 502 | return true; |
| 503 | } |
| 504 | |
| 505 | /** |
| 506 | * ResourceLoaderGetConfigVars hook handler |
| 507 | * This should be used for variables which vary with the html |
| 508 | * and for variables this should work cross skin |
| 509 | * @see https://www.mediawiki.org/wiki/Manual:Hooks/ResourceLoaderGetConfigVars |
| 510 | * |
| 511 | * @param array &$vars |
| 512 | * @param string $skin |
| 513 | * @param Config $config |
| 514 | */ |
| 515 | public function onResourceLoaderGetConfigVars( array &$vars, $skin, Config $config ): void { |
| 516 | $vars += [ |
| 517 | 'wgCirrusSearchFeedbackLink' => $config->get( 'CirrusSearchFeedbackLink' ), |
| 518 | ]; |
| 519 | } |
| 520 | |
| 521 | /** |
| 522 | * @return SearchConfig |
| 523 | */ |
| 524 | private static function getConfig() { |
| 525 | // @phan-suppress-next-line PhanTypeMismatchReturnSuperType |
| 526 | return MediaWikiServices::getInstance() |
| 527 | ->getConfigFactory() |
| 528 | ->makeConfig( 'CirrusSearch' ); |
| 529 | } |
| 530 | |
| 531 | /** |
| 532 | * @return Connection |
| 533 | */ |
| 534 | private static function getConnection() { |
| 535 | return new Connection( self::getConfig() ); |
| 536 | } |
| 537 | |
| 538 | /** |
| 539 | * Add $wgCirrusSearchInterwikiProv to external results. |
| 540 | * @param Title &$title |
| 541 | * @param string|HtmlArmor|null &$text |
| 542 | * @param SearchResult $result |
| 543 | * @param array $terms |
| 544 | * @param SpecialSearch $page |
| 545 | * @param string[] &$query |
| 546 | * @param string[] &$attributes |
| 547 | */ |
| 548 | public function onShowSearchHitTitle( &$title, &$text, $result, $terms, $page, &$query, &$attributes ) { |
| 549 | global $wgCirrusSearchInterwikiProv; |
| 550 | if ( $wgCirrusSearchInterwikiProv && $title->isExternal() ) { |
| 551 | $query["wprov"] = $wgCirrusSearchInterwikiProv; |
| 552 | } |
| 553 | } |
| 554 | |
| 555 | /** |
| 556 | * @param ApiBase $module |
| 557 | */ |
| 558 | public function onAPIAfterExecute( $module ) { |
| 559 | if ( !ElasticsearchIntermediary::hasQueryLogs() ) { |
| 560 | return; |
| 561 | } |
| 562 | $response = $module->getContext()->getRequest()->response(); |
| 563 | $response->header( 'X-Search-ID: ' . Util::getRequestSetToken() ); |
| 564 | if ( $module instanceof ApiOpenSearch ) { |
| 565 | $types = ElasticsearchIntermediary::getQueryTypesUsed(); |
| 566 | if ( $types ) { |
| 567 | $response->header( 'X-OpenSearch-Type: ' . implode( ',', $types ) ); |
| 568 | } |
| 569 | } |
| 570 | } |
| 571 | |
| 572 | /** |
| 573 | * @param string $term |
| 574 | * @param ISearchResultSet|null &$titleMatches |
| 575 | * @param ISearchResultSet|null &$textMatches |
| 576 | */ |
| 577 | public function onSpecialSearchResults( $term, &$titleMatches, &$textMatches ) { |
| 578 | $context = RequestContext::getMain(); |
| 579 | $out = $context->getOutput(); |
| 580 | |
| 581 | $out->addModules( 'ext.cirrus.serp' ); |
| 582 | |
| 583 | $jsVars = [ |
| 584 | 'wgCirrusSearchRequestSetToken' => Util::getRequestSetToken(), |
| 585 | ]; |
| 586 | // In theory UserTesting should always have been activated by now, but if |
| 587 | // somehow it wasn't we don't want to activate it now at the end of the request |
| 588 | // and report incorrect data. |
| 589 | if ( UserTestingStatus::hasInstance() ) { |
| 590 | $ut = UserTestingStatus::getInstance(); |
| 591 | if ( $ut->isActive() ) { |
| 592 | $trigger = $ut->getTrigger(); |
| 593 | $jsVars['wgCirrusSearchActiveUserTest'] = $trigger; |
| 594 | // bc for first deployment, some users will still have old js. |
| 595 | // Should be removed in following deployment. |
| 596 | $jsVars['wgCirrusSearchBackendUserTests'] = $trigger ? [ $trigger ] : []; |
| 597 | } |
| 598 | } |
| 599 | $out->addJsConfigVars( $jsVars ); |
| 600 | |
| 601 | // This ignores interwiki results for now...not sure what do do with those |
| 602 | ElasticsearchIntermediary::setResultPages( [ |
| 603 | $titleMatches, |
| 604 | $textMatches |
| 605 | ] ); |
| 606 | } |
| 607 | |
| 608 | /** |
| 609 | * @param array &$extraStats |
| 610 | * @return void |
| 611 | */ |
| 612 | private static function addWordCount( array &$extraStats ): void { |
| 613 | $search = new CirrusSearch(); |
| 614 | |
| 615 | $status = $search->countContentWords(); |
| 616 | if ( !$status->isOK() ) { |
| 617 | return; |
| 618 | } |
| 619 | $wordCount = $status->getValue(); |
| 620 | if ( $wordCount !== null ) { |
| 621 | $extraStats['cirrussearch-article-words'] = $wordCount; |
| 622 | } |
| 623 | } |
| 624 | |
| 625 | /** @inheritDoc */ |
| 626 | public function onGetPreferences( $user, &$prefs ) { |
| 627 | $search = new CirrusSearch(); |
| 628 | $profiles = $search->getProfiles( \SearchEngine::COMPLETION_PROFILE_TYPE, $user ); |
| 629 | if ( !$profiles ) { |
| 630 | return; |
| 631 | } |
| 632 | $options = self::autoCompleteOptionsForPreferences( $profiles ); |
| 633 | if ( !$options ) { |
| 634 | return; |
| 635 | } |
| 636 | $prefs['cirrussearch-pref-completion-profile'] = [ |
| 637 | 'type' => 'radio', |
| 638 | 'section' => 'searchoptions/completion', |
| 639 | 'options' => $options, |
| 640 | 'label-message' => 'cirrussearch-pref-completion-profile-help', |
| 641 | ]; |
| 642 | } |
| 643 | |
| 644 | /** |
| 645 | * @param array[] $profiles |
| 646 | * @return string[] |
| 647 | */ |
| 648 | private static function autoCompleteOptionsForPreferences( array $profiles ): array { |
| 649 | $available = array_column( $profiles, 'name' ); |
| 650 | // Order in which we propose comp suggest profiles |
| 651 | $preferredOrder = [ |
| 652 | 'fuzzy', |
| 653 | 'fuzzy-subphrases', |
| 654 | 'strict', |
| 655 | 'normal', |
| 656 | 'normal-subphrases', |
| 657 | 'classic' |
| 658 | ]; |
| 659 | $messages = []; |
| 660 | foreach ( $preferredOrder as $name ) { |
| 661 | if ( in_array( $name, $available ) ) { |
| 662 | $display = wfMessage( "cirrussearch-completion-profile-$name-pref-name" )->escaped() . |
| 663 | new \OOUI\LabelWidget( [ |
| 664 | 'classes' => [ 'oo-ui-inline-help' ], |
| 665 | 'label' => wfMessage( "cirrussearch-completion-profile-$name-pref-desc" )->text() |
| 666 | ] ); |
| 667 | $messages[$display] = $name; |
| 668 | } |
| 669 | } |
| 670 | // At least 2 choices are required to provide the user a choice |
| 671 | return count( $messages ) >= 2 ? $messages : []; |
| 672 | } |
| 673 | |
| 674 | /** @inheritDoc */ |
| 675 | public function onUserGetDefaultOptions( &$defaultOptions ) { |
| 676 | $defaultOptions['cirrussearch-pref-completion-profile'] = |
| 677 | $this->configFactory->makeConfig( 'CirrusSearch' )->get( 'CirrusSearchCompletionSettings' ); |
| 678 | } |
| 679 | |
| 680 | /** @inheritDoc */ |
| 681 | public function onSpecialStatsAddExtra( &$extraStats, $context ) { |
| 682 | self::addWordCount( $extraStats ); |
| 683 | } |
| 684 | |
| 685 | /** @inheritDoc */ |
| 686 | public function onAPIQuerySiteInfoStatisticsInfo( &$extraStats ) { |
| 687 | self::addWordCount( $extraStats ); |
| 688 | } |
| 689 | } |